On Wed Feb 5 18:49:50 2025 +0000, Shmerl wrote:
Is it practical to expect every client to behave correctly though? I.e. if something is bugged, you don't want your client to be dragged down becasue you expected every other client to be perfect? So defensive measures make sense since bugs always can happen. I.e. it's more reasonable to expect compositor to be compliant / perfect (since there are more limited number of them and they are relatively well known), but I don't think you can make such assumption about some random clients.
Which is why I think a properly designed desktop communication protocol should do something better than just connect clients together through basic IPC to then let them figure out how to communicate, putting the risk and burden of handling badly implemented clients on each one of them.
Every thing I discover with these Wayland review feels backwards, all the freedom we need to properly implement Win32 API is unavailable, and every time I expect something to be slightly better abstracted than before to hide or remove the complexity and implementation burden out of clients, it is not.
I'll look into this just to make sure it works, and if we have to do this kind of thing so be it, but honestly, if this is the future of Linux desktop, it is very depressing to say the least.